Verification Environment Checking via Requirement Similarity Feedback

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing verification processes in digital hardware design suffer from errors due to incorrect or incomplete interpretation of natural language artefacts into machine language descriptions, leading to incomplete or incorrect verification environments.

Innovation Solution

A method and device that quantifies the similarity between initial natural language descriptions and extracted machine language descriptions by comparing extracted requirements, features, and verification items, ensuring accurate translation and reducing errors through automated verification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If natural language descriptions are manually interpreted into machine language descriptions by engineers, then the verification environment can be created, but interpretation errors occur leading to incorrect or incomplete verification

Engineering Contradiction:

Improveverification accuracyVSAvoidinterpretation correctness

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The patent implements a feedback mechanism where the system automatically compares the original natural language requirements with the extracted natural language descriptions from the generated machine language code. This comparison loop identifies interpretation errors and allows for correction, ensuring that the verification environment accurately reflects the original requirements without manual intervention errors.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system creates a natural language copy of the machine language description by extracting and translating code back into natural language requirements. This copied representation is then compared with the original requirements to verify accuracy, eliminating the need for manual interpretation while preserving the original intent through automated linguistic transformation.

Inventive Principle:

Principle #26Copying

2Reliability

If automated translation from natural language to machine language is implemented, then interpretation errors are reduced, but verification of translation accuracy becomes more complex

Engineering Contradiction:

Improvetranslation accuracyVSAvoidverification system complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

Instead of verifying machine language code against natural language requirements directly, the system inverts the approach by translating the machine language back into natural language and comparing the translated text with the original requirements. This inversion simplifies the verification process by using natural language comparison techniques rather than complex code analysis.

Inventive Principle:

Principle #13The other way round (Inversion)

3Reliability

If manual review of machine language descriptions is performed to ensure compliance with natural language requirements, then accuracy is maintained, but productivity decreases

Engineering Contradiction:

Improvecompliance accuracyVSAvoidverification throughput

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system performs self-verification by automatically comparing its own generated machine language descriptions (translated back to natural language) against the original requirements. This self-service mechanism eliminates the need for manual review while maintaining high accuracy, as the system autonomously identifies and corrects its own interpretation errors.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent introduces natural language as an intermediary layer between the original requirements and the machine language code. By translating code back to natural language for comparison purposes, the system creates a mediating representation that simplifies verification while maintaining accuracy, avoiding both manual review and direct complex code analysis.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentEP4703958A1Method to check a machine language description of a verification environment, verification device, computer program product and computer-readable storage medium

Publication Date: 2026.03.04 SIEMENS AG

Figure 1
Figure 2

AI summary

The invention relates to a method to check a machine language description of a verification environment, the method comprising the following steps performed by a verification device (10): - Receiving from a control device an initial set of natural language descriptions comprising initial requirements, initial features and initial verification items related to the verification environment; - Receiving from the control device the machine language description of the verification environment; Analyzing the machine language description of the verification environment in a predefined extraction procedure to generate an extracted set of natural language descriptions comprising extracted requirements, extracted features and extracted verification items based on the machine language description of the verification environment; Comparing the extracted set of natural language descriptions with the initial set of natural language descriptions to evaluate a degree of similarity between the sets, using a language interpretation module of the verification device (10) in a comparison procedure.